Academic collaboration has become the dominant mode of knowledge production in artificial intelligence research. Analysis of over 464,000 AI articles published between 1991 and 2020 reveals that only 8.6% are single-authored, while nearly half (48.9%) have two or three authors, and another 34.9% have four to six authors. Moreover, more than half (53.8%) of these publications involve researchers from multiple organizations, highlighting the cross-institutional nature of modern AI research. [15] These collaborative relationships form specialized academic networks known as co-authorship networks, where authors are represented as vertices and collaborations between authors as edges. Understanding these networks has become increasingly important in the academic landscape.
Rachel So. AI Co-authorship in Academic Publishing. Project Rachel, April 2025. https://doi.org/10.71775/kth.206ra-axe33